In this case, they were correct. We had a lengthy discussion on the post of Safire's column, searched high and low through the Homeland Security bill, and could only find some funding to prod companies to develop the technology required for such a database. But IMO the administrative problems for such a database are far more significant. You will literally have hundreds of thousands of data feeds into the system, if not millions. You will have to analyze each feed, load it, parse it, perform data hygiene on it, match it, dedupe it, and then build models from it. That IMO is a far more daunting task than anyone in government cares to acknowledge, and it consistent with the general approach of the feds to attempt to throw technology at administration/bureaucratic problems, with predictable results.
Could be I read it wrong, but that's what I saw.